{"id":"W4401609584","doi":"10.1109/icasspw62465.2024.10626010","title":"Characterizing the Temporal Dynamics of Universal Speech Representations for Generalizable Deepfake Detection","year":2024,"lang":"en","type":"article","venue":"","topic":"Speech Recognition and Synthesis","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"ca_institutions":"Institut National de la Recherche Scientifique","funders":"","keywords":"Dynamics (music); Computer science; Speech recognition; Artificial intelligence; Natural language processing; Acoustics; Physics","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001807894,0.001280027,0.0008176402,0.001194878,0.0004128315,0.0009920318,0.0006959971,0.0009261867,0.001735873],"category_scores_gemma":[0.007143133,0.0003520325,0.0005383907,0.0006111396,0.000762019,0.00207737,0.001676067,0.002092426,0.00113102],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005446943,"about_ca_system_score_gemma":0.0007260588,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001726826,"about_ca_topic_score_gemma":0.003408534,"domain_scores_codex":[0.9992073,0.000179079,0.0000486866,0.0002935321,0.0001533242,0.0001180727],"domain_scores_gemma":[0.9971309,0.001399194,0.0002651183,0.0007633053,0.0003064477,0.000135044],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.0007040378,0.0002953244,0.01434832,0.0002665706,0.0001634333,0.0003358491,0.0004133257,0.164672,0.07415678,0.009072178,0.006515149,0.729057],"study_design_scores_gemma":[0.00001451622,0.000136822,0.004811571,0.00003274611,0.00003702088,0.0002733009,0.00009826912,0.9646564,0.01857878,0.009057631,0.00227467,0.00002835081],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2259035,0.001409038,0.7630302,0.0005520065,0.00018709,0.0001235685,0.0009208548,0.004237306,0.003636348],"genre_scores_gemma":[0.9158699,0.0004093563,0.07869773,0.0001658008,0.00007148906,0.00008385471,0.001738158,0.0002978944,0.002665812],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.001807894,"threshold_uncertainty_score":0.009561181,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0281430471174211,"score_gpt":0.2683847684010593,"score_spread":0.2402417212836382,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}